substances

 31

DEHP exposure (Lange et al. 2021 and Deliverable D 5.2 in HBM4EU). There are also limited data for the development of a PBK* model for occupational exposure via inhalation and skin absorption.

3.2.2. Main uncertainties related to the derivation of an OBL* for human toxicity datarich substances

The main uncertainties related to the derivation of OBLs* that are directly based on correlations between biomarker and health effects (method 1), are the strength of the evidence for the correlation, possible confounding factors, and biases. An experienced risk assessor needs to evaluate the consistencies of the epidemiological studies for the identified health effect as well as the biological plausibility of this effect and available animal or other mechanistic or kinetic data supporting these effects in humans.

OBLs*derived indirectly using established correlations between air and biomarker levels are associated with the existing OELs* uncertainties. These are related to: 1) the underlying toxicological and epidemiological data used to derive dose-response from external exposures, and 2) the correlations between external and internal exposures. The most significant uncertainties are often related to the human relevance of available animal data and extrapolations from animal to human. This is especially applicable to animal toxicity data used to derive NOAELs*, LOAELs*, and BMDs*. Although these uncertainties can be covered by assessment factors (which are often only “default” factors, not based on real data on the level of the uncertainty, see chapter 3.3.5, they do not remove the underlying uncertainty related to human relevance. If high assessment factors are needed to account for the uncertainties in toxicological (dose-response) data, small uncertainties, such as extrapolations of external exposure to internal biomarker level become less important. As discussed above in the case of Cr(VI), when evaluating the reliability of the established correlations, attention should be paid to: 1) the range of air and urine concentrations covered by the measured data, 2) the number of data points, 3) strength of the correlation, and if there are several studies for the same substance, 4) coherence between studies. Correlations obtained from controlled volunteer studies may be more reliable compared to the correlations obtained from the workplace with their inherent confounding factors.

Table 5. Summary of main uncertainties in OBL* derivations

Approach

1. Derivation of health-based biomonitoring values directly based on the data on correlations between biomarker and health effects.

2. Derivation of biomonitoring values indirectly using established correlations between air and biomarker levels and existing OELs* or OELV*.

Main uncertainties and reasons for variation

General uncertainties related to epidemiological data, i.e., confounding factors related to the data, biases. Consistency of the data Biological plausibility 1) the underlying toxicological and epidemiological data used to derive dose-responses from external exposures. High uncertainties related to the toxicological database, the greater the assessment factors.

2) the correlations between external and internal levels data. When evaluating the correlation data, attention should be paid to:

the strength of the correlation, availability of multiple studies showing correlation, consistency of the correlations between the studies and possible reasons for the variation (e.g., the extent of skin exposure may explain variation in some cases), exposure ranges cover the interested concentration (necessary to extrapolate to much lower levels?) In case of metals and metalloids also the information on speciation of metal compound and health relevant fraction, e.g. percentage of the alveolar fraction with a specific diameter is needed for assessment of results

OCCUPATIONAL BIOMONITORING GUIDANCE DOCUMENT © OECD 2022